prune

PriorThresholdPruner.prune(state)

Removes components from the mixture whose weights are below the threshold.

This method inspects state.curr_mixture and removes any component whose weight is less than threshold. The pruning logic ensures that at least one component always remains in the model. The weights of the remaining components are automatically re-normalized by the rework_pysatl_mpest.core.MixtureModel.

Parameters:

state (PipelineState) – The current state of the pipeline, containing the mixture model to be pruned.

Returns:

The updated pipeline state. If components were removed, state.curr_mixture will be a new, smaller rework_pysatl_mpest.core.MixtureModel instance. Otherwise, the original state is returned.

Return type:

PipelineState